Skip to main content

VIP Features

This document offers a comprehensive overview of the VIP Services API functions designed for advanced video processing, analytics, and avatar-based content generation.

Video Summary Llava

Utilize the video_process_llava method to generate a summary for a video.

ParameterTypeRequiredDescription
videoUrlStringYesS3 URL of the input video file
instructionStringYesInstruction/query to summarize

# Initialize the SDK

videoSummary = sdk.video_process_llava(

videoUrl="https://aws.com/s3-videos/video.mp4",

instruction="Give me the summary of this video",

)

print(videoSummary)

Video Chat

Use the video_chat method to inquire about the source video (Source Type must be video).

ParameterTypeRequiredDescription
source_idIntegerYesID of the video-type source
questionStringYesQuestion to ask about the video

# Initialize the SDK

video_chat = sdk.video_chat(source_id=113, question="What happened in this video?")

print(video_chat)

Employ the search_in_video method to locate objects in the video (Source Type must be video).

ParameterTypeRequiredDescription
source_idIntegerYesID of the video-type source
search_textStringYesObject to search in the video

# Initialize the SDK

search_in_video = sdk.search_in_video(source_id=113, search_text="Red Car")

print(search_in_video)

Fetching Raw Analytics

Use the get_raw_analytics method to retrieve raw analytics for a source.

  • For video source type, only source_id is required.
  • For camera source type, it requires source_id, start_time, and end_time to obtain raw analytics.
  • Time Format: "2025-05-18T15:32:38.105Z"
ParameterTypeRequiredDescription
source_idIntegerYesID of the source
start_timeStringYes (camera) No (video)Start time in ISO format
end_timeStringYes (camera) No (video)End time in ISO format

# Initialize the SDK

getRawAnalytics = sdk.get_raw_analytics(source_id=2, start_time=startTime, end_time=endTime)

print(getRawAnalytics)

Fetching Analytics Report

Utilize the get_analytics_report method to obtain raw analytics in CSV format for a source.

  • For video source type, only source_id is required.
  • For camera source type, it requires source_id, start_time, and end_time to obtain raw analytics.
  • Time Format: "2025-05-18T15:32:38.105Z"
  • This method saves the CSV file.
ParameterTypeRequiredDescription
source_idIntegerYesID of the source
start_timeStringYes (camera) No (video)Start time in ISO format
end_timeStringYes (camera) No (video)End time in ISO format

# Initialize the SDK

report = sdk.get_analytics_report(source_id=2, start_time=startTime, end_time=endTime)

print(report)

Fetching Analytics Report List

Use the get_analytics_report_list method to retrieve raw analytics in CSV format for multiple sources.

  • It requires source_ids in a list, start_time, and end_time to obtain raw analytics.
  • Time Format: "2025-05-18T15:32:38.105Z"
  • This method saves the CSV file.
ParameterTypeRequiredDescription
source_idListYesIDs of the sources
start_timeStringYes (camera) No (video)Start time in ISO format
end_timeStringYes (camera) No (video)End time in ISO format

# Initialize the SDK

reportsList = sdk.get_analytics_report_list(source_ids=[125, 239], start_time=startTime, end_time=endTime)

print(reportsList)

Video Library

Use the video_library method to access event videos associated with an analytics ID or source.

ParameterTypeDefaultDescription
analytic_idsListRequiredList of analytics IDs to fetch related videos
zone_idIntegerOptionalFilter results by zone ID
source_idIntegerOptionalFilter results by source ID
page_numberInteger0Pagination - page number
page_sizeInteger6Pagination - items per page

# Initialize the SDK

video_lib = sdk.video_library(

analytic_ids=[42],

source_id=139

)

print(video_lib)

Fetching Avatars

Fetching Avatars

Utilize the get_avatar_list method to retrieve a list of avatars for generating agent videos.

{get_agents.map((agent, idx) => (
<div key={idx}>
<p>Name: {agent.name}, Url: {agent.avatarUri}</p>
</div>
))}

### Generate Videos
Employ the generate\_video method to create an avatar video based on the specified text.

|**Parameter**|**Type**|**Required**|**Description**|
| :-: | :-: | :-: | :-: |
|name|String|Yes|Name/title of the video|
|gender|String|Yes|Avatar gender: "male" or "female"|
|language|String|Yes|Language for speech (e.g., "English", "Hindi")|
|text|String|Yes|Context for the avatar to speak|
|avatar\_url|String|Yes|URL to the avatar image/video|

\# Initialize the SDK

gevVideo = sdk.generate\_video(

` `name="Test SDK",

` `gender="female",

` `language="Korean",

` `text="Hello, this is a test.",

` `avatar\_url="https://cdn-dev.eizen.ai/6b9fb1f5-a859-4883-b159-941908bbe886.png"

)

print(gevVideo)
### Generated Video Status
Use the check\_generation\_video\_status method to verify the status of the avatar video generation.

|**Parameter**|**Type**|**Required**|**Description**|
| :-: | :-: | :-: | :-: |
|video\_id|Integer|Yes|ID of the generating video|

\# Initialize the SDK

checkStatus = sdk.check\_generation\_video\_status(video\_id=124)

print(checkStatus)
#### Notes
- Ensure that the SDK is initialized properly with a valid User token before invoking these methods.
- All returned data is typically formatted as a dictionary or list.